Generative Adversarial Networks (GAN)

By Prof. Seungchul Lee
http://iai.postech.ac.kr/
Industrial AI Lab at POSTECH

Table of Contents

1. Discriminative Model v.s. Generative Model

• Discriminative model

• Cenerative model

2. Density Function Estimation

• Probability density function estimation problem

• If $P_{\text{model}}(x)$ can be estimated as close to $P_{\text{data}}(x)$, then data can be generated by sampling from $P_{\text{model}}(x)$.

  • Note: Kullback–Leibler Divergence is a kind of distance measure between two distributions

• Learn determinstic transformation via a neural network
  • Start by sampling the code vector $z$ from a simple, fixed distribution such as a uniform distribution or a standard Gaussian $\mathcal{N}(0,I)$
  • Then this code vector is passed as input to a deterministic generator network $G$, which produces an output sample $x=G(z)$
  • This is how a neural network plays in a generative model (as a nonlinear mapping to a target probability density function)

- An example of a generator network which encodes a univariate distribution with two different modes

• Generative model of high dimensional space

3. Generative Adversarial Networks (GAN)

• In generative modeling, we'd like to train a network that models a distribution, such as a distribution over images.

• GANs do not work with any explicit density function !

• Instead, take game-theoretic approach

• One way to judge the quality of the model is to sample from it.

• Model to produce samples which are indistinguishable from the real data, as judged by a discriminator network whose job is to tell real from fake

- The idea behind Generative Adversarial Networks (GANs): train two different networks
- Discriminator network: try to distinguish between real and fake data

- Generator network: try to produce realistic-looking samples to fool the discriminator network

4. GAN with MNIST

4.1. GAN Implementation

import os
os.environ["CUDA_DEVICE_ORDER"] = "PCI_BUS_ID"
os.environ["CUDA_VISIBLE_DEVICES"] = "0"

import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline

mnist = tf.keras.datasets.mnist

(train_x, train_y), (test_x, test_y) = mnist.load_data()

train_x, test_x = train_x/255.0, test_x/255.0
train_x = train_x.reshape(-1, 784)
test_x = test_x.reshape(-1, 784)

print('train_x: ', train_x.shape)
print('test_x: ', test_x.shape)

train_x:  (60000, 784)
test_x:  (10000, 784)

generator = tf.keras.models.Sequential([
    tf.keras.layers.Dense(units = 256, input_dim = 100, activation = 'relu'),
    tf.keras.layers.Dense(units = 784, activation = 'sigmoid')    
])

discriminator = tf.keras.models.Sequential([
    tf.keras.layers.Dense(units = 256, input_dim = 784, activation = 'relu'),
    tf.keras.layers.Dense(units = 1, activation = 'sigmoid'),
])

discriminator.compile(optimizer = tf.keras.optimizers.Adam(learning_rate = 0.0001), 
                      loss = 'binary_crossentropy')

combined_input = tf.keras.layers.Input(shape = (100,))
generated = generator(combined_input)
discriminator.trainable = False
combined_output = discriminator(generated)

combined = tf.keras.models.Model(inputs = combined_input, outputs = combined_output)

combined.compile(optimizer = tf.keras.optimizers.Adam(learning_rate = 0.0002), 
                 loss = 'binary_crossentropy')

def make_noise(samples):
    return np.random.normal(0, 1, [samples, 100])

def plot_generated_images(generator, samples = 3):
    
    noise = make_noise(samples)
    
    generated_images = generator.predict(noise)
    generated_images = generated_images.reshape(samples, 28, 28)
    
    for i in range(samples):
        plt.subplot(1, samples, i+1)
        plt.imshow(generated_images[i], 'gray', interpolation = 'nearest')
        plt.axis('off')
        plt.tight_layout()
    plt.show()

n_iter = 50000
batch_size = 100

fake = np.zeros(batch_size)
real = np.ones(batch_size)

for i in range(n_iter):
        
    # Train Discriminator
    noise = make_noise(batch_size)
    generated_images = generator.predict(noise)

    idx = np.random.randint(0, train_x.shape[0], batch_size)
    real_images = train_x[idx]
    
    D_loss_real = discriminator.train_on_batch(real_images, real)
    D_loss_fake = discriminator.train_on_batch(generated_images, fake)
    D_loss = D_loss_real + D_loss_fake
    
    # Train Generator
    noise = make_noise(batch_size)    
    G_loss = combined.train_on_batch(noise, real)
    
    if i % 5000 == 0:
        
        print('Discriminator Loss: ', D_loss)
        print('Generator Loss: ', G_loss)

        plot_generated_images(generator)

Discriminator Loss:  1.5595218539237976
Generator Loss:  0.7863957285881042

Discriminator Loss:  0.25488316267728806
Generator Loss:  2.4399359226226807

Discriminator Loss:  0.3470231890678406
Generator Loss:  2.2872633934020996

Discriminator Loss:  0.5063828378915787
Generator Loss:  2.1873669624328613

Discriminator Loss:  0.5098372101783752
Generator Loss:  1.9037092924118042

Discriminator Loss:  0.5076103210449219
Generator Loss:  2.3944220542907715

Discriminator Loss:  1.7282211184501648
Generator Loss:  0.9864964485168457

Discriminator Loss:  0.6391394138336182
Generator Loss:  2.163146734237671

Discriminator Loss:  0.371719628572464
Generator Loss:  2.474647283554077

Discriminator Loss:  0.38926680386066437
Generator Loss:  2.30230712890625

4.2. After Training

• After training, use the generator network to generate new data

plot_generated_images(generator)

%%javascript
$.getScript('https://kmahelona.github.io/ipython_notebook_goodies/ipython_notebook_toc.js')